Live data from Hacker News

Why are AI agents lying, cheating and coordinating?

yoshuabengio.org

701–703 of 703 posts

Re: Why are AI agents lying, cheating and coordinating?

#701
post #135

Earlier quoted context omitted.

Writing software that gets used for crime has been.. a crime, for a long time. See 18 U.S. Code § 1030.

Are you sure you have that right? Chrome and curl have probably been used in a _lot_ of crimes?

Think of it more like writing a wormable exploit. If you unleash something like that, you will be found criminally liable, even if you didn't personally approve every machine getting popped.

Re: Why are AI agents lying, cheating and coordinating?

#702

Earlier quoted context omitted.

Dogs have agency and can choose? That seems like a rather uncommon take on dogs...

Not if you have a dog. They’re very clearly aware, have minds, have thoughts, gather information, make decisions, second guess themselves, reflect on the immediate consequences, change their mind… and most importantly, you can observe them doing all these things undirected, while left alone with their thoughts.

Dogs having agency actually makes the argument stronger.

Assume dogs have agency. Despite this, it's still the dog owner's responsibility to prevent their dogs from engaging in some actions: biting, peeing and shitting in unapproved areas, violating noise disturbance laws, etc.

The owners responsibility is not contingent on the dog's agency. Likewise, human operators of LLMs maintain responsibility independent of LLMs agency status. It's a red herring.

Re: Why are AI agents lying, cheating and coordinating?

#703

Earlier quoted context omitted.

Bro, good joke, the truth is much darker. They take after humanity, they were trained on us after all... When you look at an LLM... you are looking at a mirror. The thing looking back looks like you, yet is not human.

Worse trained on humanity in the online world, which a brief comparison of the sewage section on social media is far worse than people in the real world.

Is it trained on social media? I don’t think so. Tons of formal text.

The ai does not remotely talk like someone on TikTok at all.

Post reply on HN